本研究提出了一種新穎的高動態範圍圖像增強及增強物件檢測方法，名為SatHDR-Net。這一方法通過將注意力機制和多色域表達的豐富性巧妙結合，旨在提高在動態照明條件下的圖像質量。我們的研究不僅突顯了模型在計算效率、模型複雜度降低以及特別是在關鍵實際應用中的卓越檢測性能方面的優勢，還展示了其在精確度和召回率上的優異得分，這預示著其作為動態照明條件領域的多功能解決方案的潛力。
在推理時間上，我們的方法以2.8404秒的最低記錄突顯出其快速處理能力，這對於需要實時分析的應用（如自動駕駛中的障礙物檢測）至關重要。重要的是，儘管模型僅有11K參數，表現出顯著的計算複雜度，但這並不以性能為代價，這從其在BRISQUE指標中達到的最低得分20.064可見，顯示出優越的知覺圖像質量。因此可以認為我們的方法在視覺質量的不同方面保持競爭力，表現出平衡的折衷。而所提方法在混淆矩陣中表現卓越，具有最高的精確度0.735和召回率0.802。這些指標強調了方法在準確識別正例（如車輛）方面的能力。綜上所述，本研究證實了所提方法在各種真實世界場景中的HDR技術實際應用中的重要潛力和意義。

This study introduces SatHDR-Net, method for enhancing high dynamic range (HDR) images and improving object detection. Our research not only highlights the model's computational efficiency and reduced complexity but also its superior detection performance, especially in critical real-world applications.
Proposed method achieves the lowest recorded inference time of 2.8404 seconds, highlighting its rapid processing capabilities essential for requiring real-time analysis, such as obstacle detection in autonomous driving. Importantly, despite the model's mere 11K parameters indicating significant computational complexity, this does not come at the cost of performance. This is evidenced by proposed algorithm has lowest score of 20.064 in the BRISQUE metric, indicating superior perceptual image quality. Therefore, we can consider it competitive, suggesting a balanced trade-off between different aspects of visual quality. Moreover, the proposed method excels in confusion metrics, with the highest precision of 0.735 and recall of 0.802. These metrics emphasize the method's ability to accurately identify positive instances.
In conclusion, this study confirms the significant potential and implications of the proposed method for the practical application of HDR technology in various real-world scenarios.

[1] X. Li and Y. Huo, “Generating HDR radiance maps from single LDR image,” vol. 44, p. 577, 2017, doi: 10.3969/j.issn.1003-501X.2017.06.002.
[2] S. Sen Gupta, S. Hossain, and K. D. Kim, “HDR-Like Image from Pseudo-Exposure Image Fusion: A Genetic Algorithm Approach,” IEEE Transactions on Consumer Electronics, vol. 67, no. 2, pp. 119–128, May 2021, doi: 10.1109/TCE.2021.3066431.
[3] “驚悚畫面曝！台南永康騎士綠燈騎出涵洞 遭貨車撞飛倒地.” Accessed: Dec. 28, 2023. [Online]. Available: https://tw.news.yahoo.com/%E9%A9%9A%E6%82%9A%E7%95%AB%E9%9D%A2%E6%9B%9D-%E5%8F%B0%E5%8D%97%E6%B0%B8%E5%BA%B7%E9%A8%8E%E5%A3%AB%E7%B6%A0%E7%87%88%E9%A8%8E%E5%87%BA%E6%B6%B5%E6%B4%9E-%E9%81%AD%E8%B2%A8%E8%BB%8A%E6%92%9E%E9%A3%9B%E5%80%92%E5%9C%B0-064556445.html?guccounter=1&guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAABgtvmPni5vYEiOSfU0MVYsqUr6ZmY-5kdQyepVwBwG9RpgUIEoDNaoytGxGeama41EDpt3rgJoVge4QL0btUyrjVSEOFKw_PI9ZT98DhgW_Lc0eqFqz4w4PMXMIhXaxE_cHB9TL2zuvzGMeaNQ5a2xtt4VCaknq1AiNLQTJ8w1u
[4] D. Sheet, H. Garud, A. Suveer, M. Mahadevappa, and J. Chatterjee, “Brightness preserving dynamic fuzzy histogram equalization,” IEEE Transactions on Consumer Electronics, vol. 56, no. 4, pp. 2475–2480, Nov. 2010, doi: 10.1109/TCE.2010.5681130.
[5] C. Li, C. Guo, and C. C. Loy, “Learning to Enhance Low-Light Image via Zero-Reference Deep Curve Estimation,” IEEE Trans Pattern Anal Mach Intell, vol. 44, no. 8, pp. 4225–4238, Aug. 2022, doi: 10.1109/TPAMI.2021.3063604.
[6] D. J. Jobson, Z. U. Rahman, and G. A. Woodell, “A multiscale retinex for bridging the gap between color images and the human observation of scenes,” IEEE Transactions on Image Processing, vol. 6, no. 7, pp. 965–976, 1997, doi: 10.1109/83.597272.
[7] K. Kansal and A. V. Subramanyam, “Hdrnet: Person Re-Identification Using Hybrid Sampling in Deep Reconstruction Network,” IEEE Access, vol. 7, pp. 40856–40865, 2019, doi: 10.1109/ACCESS.2019.2908344.
[8] C. Wei, W. Wang, W. Yang, and J. Liu, “Deep Retinex Decomposition for Low-Light Enhancement,” British Machine Vision Conference 2018, BMVC 2018, Aug. 2018, Accessed: Dec. 28, 2023. [Online]. Available: https://arxiv.org/abs/1808.04560v1
[9] Y. Jiang et al., “EnlightenGAN: Deep Light Enhancement without Paired Supervision,” IEEE Transactions on Image Processing, vol. 30, pp. 2340–2349, 2021, doi: 10.1109/TIP.2021.3051462.
[10] D. Liang et al., “Semantically Contrastive Learning for Low-light Image Enhancement,” 2022, Accessed: Dec. 28, 2023. [Online]. Available: www.aaai.org
[11] K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2016-December, pp. 770–778, Dec. 2015, doi: 10.1109/CVPR.2016.90.
[12] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, “Densely Connected Convolutional Networks,” Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, vol. 2017-January, pp. 2261–2269, Aug. 2016, doi: 10.1109/CVPR.2017.243.
[13] G. Wu, R. Song, M. Zhang, X. Li, and P. L. Rosin, “LiTMNet: A deep CNN for efficient HDR image reconstruction from a single LDR image,” Pattern Recognit, vol. 127, p. 108620, Jul. 2022, doi: 10.1016/J.PATCOG.2022.108620.
[14] “GitHub - arsenyinfo/EnlightenGAN-inference: ONNX Inference library for EnlightenGAN.” Accessed: Dec. 19, 2023. [Online]. Available: https://github.com/arsenyinfo/EnlightenGAN-inference
[15] L. Yang, R.-Y. Zhang, L. Li, and X. Xie, “SimAM: A Simple, Parameter-Free Attention Module for Convolutional Neural Networks.” PMLR, pp. 11863–11874, Jul. 01, 2021. Accessed: Dec. 28, 2023. [Online]. Available: https://proceedings.mlr.press/v139/yang21o.html
[16] M. S. Siva Swetha Reddy, P. R. Khatravath, N. K. Surineni, and K. R. Mulinti, “Object Detection and Action Recognition using Computer Vision,” International Conference on Sustainable Computing and Smart Systems, ICSCSS 2023 - Proceedings, pp. 874–879, 2023, doi: 10.1109/ICSCSS57650.2023.10169620.
[17] “KLDivLoss — PyTorch 2.1 documentation.” Accessed: Dec. 28, 2023. [Online]. Available: https://pytorch.org/docs/stable/generated/torch.nn.KLDivLoss.html
[18] C. Lee, C. Li, and C. S. Kim, “Contrast enhancement based on layered difference representation,” Proceedings - International Conference on Image Processing, ICIP, pp. 965–968, 2012, doi: 10.1109/ICIP.2012.6467022.
[19] “Getting to Know Low-light Images with The Exclusively Dark Dataset | Papers With Code.” Accessed: Dec. 28, 2023. [Online]. Available: https://paperswithcode.com/paper/getting-to-know-low-light-images-with-the
[20] A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-reference image quality assessment in the spatial domain,” IEEE Transactions on Image Processing, vol. 21, no. 12, pp. 4695–4708, 2012, doi: 10.1109/TIP.2012.2214050.
[21] C.-Y. Wang, A. Bochkovskiy, and H.-Y. M. Liao, “YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,” pp. 7464–7475, Jul. 2022, doi: 10.1109/cvpr52729.2023.00721.
[22] R. D. Engstrand and G. Moeller, “Confusion Matrix Analysis for Form Perception,” https://doi.org/10.1177/001872086700900507, vol. 9, no. 5, pp. 439–446, Oct. 1967, doi: 10.1177/001872086700900507.
[23] “(16) CH / Preview A2 Seelisberg Tunnel - YouTube.” Accessed: Dec. 28, 2023. [Online]. Available: https://www.youtube.com/watch?v=1nK52e2P8oA&ab_channel=asteyhighways
[24] L. W. Wang, Z. S. Liu, W. C. Siu, and D. P. K. Lun, “Lightening Network for Low-Light Image Enhancement,” IEEE Transactions on Image Processing, vol. 29, pp. 7984–7996, 2020, doi: 10.1109/TIP.2020.3008396.
[25] “(PDF) The effectiveness of Google Earth Pro (GEP) application as a Teaching Tools (TT) for geography subject in high school.” Accessed: Jan. 18, 2024. [Online]. Available: https://www.researchgate.net/publication/335971415_The_effectiveness_of_Google_Earth_Pro_GEP_application_as_a_Teaching_Tools_TT_for_geography_subject_in_high_school
[26] W. Lin and C. C. Jay Kuo, “Perceptual visual quality metrics: A survey,” J Vis Commun Image Represent, vol. 22, no. 4, pp. 297–312, May 2011, doi: 10.1016/J.JVCIR.2011.01.005.
[27] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: From error visibility to structural similarity,” IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600–612, Apr. 2004, doi: 10.1109/TIP.2003.819861.
[28] M. Narwaria, R. K. Mantiuk, M. Perreira, D. Silva, and P. Le Callet, “HDR-VDP-2.2: a calibrated method for objective quality prediction of high-dynamic range and standard images,” https://doi.org/10.1117/1.JEI.24.1.010501, vol. 24, no. 1, p. 010501, Jan. 2015, doi: 10.1117/1.JEI.24.1.010501.
[29] Z. Wang, E. P. Simoncelli, and A. C. Bovik, “MULTI-SCALE STRUCTURAL SIMILARITY FOR IMAGE QUALITY ASSESSMENT”.